FDA Report Collection, Analysis, and Availability of Demographic Subgroup Data for FDA-Approved Medical Products August 2013 U.S. Department of Health and Human Services Food and Drug Administration Message from Dr. Margaret Hamburg Commissioner of Food and Drugs The Food and Drug Administration Safety and Innovation Act (FDASIA) of 2012 included a requirement that FDA study the availability of data on the participation of demographic subgroups (sex, age, race, and ethnicity) in clinical trials that support applications for new drugs, biologics, and devices. Specifically, Section 907 of FDASIA directed FDA to report on the extent to which subgroups participate in such trials, whether reports of subgroup safety and effectiveness are reported to FDA in a manner consistent with FDA requirements and guidance, and whether and how safety and effectiveness data by subgroup is eventually made public. We welcome the opportunity to take a closer look at the inclusion and analysis of demographic subgroups in applications for medical products. An FDA-wide working group, tasked with the responsibility of producing the report, examined 72 product applications approved in 2011. I am pleased to announce that the report we are providing to Congress and posting on our Website today concludes that the statutes, regulations, and policies currently in place generally give product sponsors a solid framework for providing data in their applications on the inclusion and analysis of demographic subgroups. In general, sponsors are describing the demographic profiles of their clinical trial participants, and the majority of applications submitted to FDA include demographic subset analyses. We also found that FDA shares this information with the public in a variety of ways. We look forward to hearing from patients, consumers, health care practitioners, industry, and others about this report. This input, as well as the report’s key findings, especially its identified areas for improvement, will help inform our creation of an Action Plan, as required by Congress. Once we develop an Action Plan, we look forward to continuing to interact with all interested stakeholders as we implement the Action Plan. Sincerely, Margaret A. Hamburg, M.D. Commissioner of Food and Drugs 2 Table of Contents Executive Summary ........................................................................................................................ 4 Introduction ..................................................................................................................................... 7 Background ..................................................................................................................................... 9 Report Organization ..................................................................................................................... 10 Tools to Ensure Analysis of Demographic Information ............................................................ 11 Part 1: A. B. 1. 2. 3. 4. 5. 6. C. 1. 2. 3. D. Drugs and Biologics ........................................................................................................ 13 Introduction ...................................................................................................................... 13 Demographic Subgroup Information Contained in Applications .................................... 14 CDER – Age Composition ............................................................................................... 15 CDER – Sex Composition ............................................................................................... 17 CDER – Race Composition.............................................................................................. 18 CBER – Age Composition ............................................................................................... 21 CBER – Sex Composition ................................................................................................ 22 CBER – Race Composition .............................................................................................. 24 Subset Analyses in Applications and Public Availability of Data.................................... 26 CDER−Public Availability of Age Subset Data .............................................................. 27 CDER−Public Availability of Sex Subset Analysis......................................................... 29 CDER−Public Availability of Race Subset Analysis....................................................... 30 Summary of Findings for Part 1: Drugs and Biologics ................................................... 34 Part 2: Medical Devices ............................................................................................................... 35 A. Introduction ...................................................................................................................... 35 B. Demographic Subgroup Information Contained in Applications .................................... 37 1. Age Composition.............................................................................................................. 37 2. Sex Composition .............................................................................................................. 39 3. Ethnic and Racial Composition ........................................................................................ 41 C. Multiple Pivotal Studies ................................................................................................... 44 1. Age Composition.............................................................................................................. 45 2. Sex Composition .............................................................................................................. 46 3. Ethnic and Racial Composition ........................................................................................ 47 D. Subset Analysis in Applications and Availability of Summary Information .................... 50 E. Summary of Findings for Part 2: Medical Devices.......................................................... 52 FDA Tools for Communicating Demographic Information ..................................................... 53 Summary and Conclusions .......................................................................................................... 58 Appendix 1: Tools for Ensuring Demographic Data Analysis.................................................. 62 Appendix 2: Products Approved in CDER, Indications, and Abbreviations......................... 74 Appendix 3: U.S. Census Bureau 2011 “People Quick Facts” ................................................. 79 Appendix 4: Indications for Medical Device PMAs.................................................................. 80 Appendix 5: Examples of FDA Publications on Related Topics.............................................. 87 3 Executive Summary For many decades, U.S. governmental initiatives have sought to identify the best ways to advance the three ethical principles that underlie the conduct of biomedical research: respect for persons, beneficence and justice (which asks who ought to receive the benefits of research and bear its burdens). 1 Consistent with the principle of justice, FDA has a variety of statutory, regulatory, and policy-related tools that provide a framework for guiding medical product sponsors in the inclusion and analysis of demographic subgroups in clinical trials. However, scientific advances in understanding the specific genetic variables underlying disease and response to treatment are increasingly becoming the focus of modern medical product development as we move toward the ultimate goal of tailoring treatments to the individual, or class of individuals, through personalized medicine. Thus, the broad, self-identified demographic subgroup categories used today may not adequately capture the complexity underlying responses to medical treatments. Nonetheless, it remains important that clinical trials include diverse populations, whenever possible and appropriate. Last year Congress, in Section 907 of the Food and Drug Administration Safety and Innovation Act of 2012 (FDASIA), directed FDA to produce a report that took a closer look at the inclusion and analysis of demographic subgroups. Specifically, Congress asked FDA to consider four key topic areas: tools to ensure submission of demographic information, subset analysis, demographic subgroup participation in clinical trials submitted to the FDA in support of product applications, and communication of this information to health care professionals and the American public. To comply with that request, an FDA working group evaluated 72 applications approved during 2011 for new molecular entity2 drug products, original biologics, and Class III devices (premarket approval). Their key findings, organized by topic area, are as follows: Tools to ensure submission of demographic information • Although there is some variation by product area, FDA’s statutory and regulatory requirements, guidances, policies, and procedures generally inform sponsors about including tabulations of the demographic data on clinical trial participants and demographic subset analyses in their medical product applications (see Appendix 1). 1 Ethical Principles and Guidelines for the Protection of Human Subjects of Research (The Belmont Report). Available at www.hhs.gov/ohrp/humansubjects/guidance/belmont.html. Accessed July 30, 2013. 2 A new molecular entity is an active ingredient that has never before been marketed in the United States in any form. Available at: http://www.fda.gov/drugs/informationondrugs/ucm079436.htm Accessed: May 15, 2013. 4 • Similarly, tools (e.g., application review templates and FDA standard operating policies and procedures) guide regulatory review staff in the assessment of marketing applications to ensure that demographic data and subset analyses are included in the information FDA uses in its review and approval processes. Extent of demographic subset analyses • The extent to which demographic subset data were analyzed varied across medical product types (drugs, biologics, and devices). Applications for drugs and biologics addressed subset analyses by sex, race, and age − that is, the applications mentioned demographic subsets in some way. The majority of the device applications contained a subset analysis for age and sex, with a lower percentage of applications containing a subset analysis for race or ethnicity. Inclusion did not necessarily mean that the data on patient subgroups was sufficient for meaningful analysis or to detect relevant subgroup effects. In some of the applications reviewed for this report, the results of the subgroup analyses were limited by low sample size. Extent of demographic subgroup representation in clinical trials • • • All biologics, drugs, and the majority of the medical device applications reviewed for this report provided the composition of clinical study participants by age, race, and sex. Participants’ sex was the most consistently reported in the medical product applications. For approved drugs and biologics, the extent to which patients were represented in clinical trials by age and sex tended to reflect the disease indication studied. For devices, patient participation by age and sex varied by product area. Although it is important to include diverse populations in clinical trials when possible and appropriate, the unique nature of medical devices means that this additional information would not always contribute to FDA’s decision making. For example, certain microbiology in-vitro diagnostic devices (IVDDs) have high overall accuracy; when this occurs, additional analyses by subpopulations would not affect clearance or approval or clinical use. Whites represented a high percentage of clinical trial study participants for biologic, drug, and medical device applications. In many cases, other racial subgroups were underrepresented. Communication of demographic subgroup information to the public • FDA’s internal policies and procedures and regulations facilitate the assessment of demographic subgroup information included in marketing applications. Moreover, following medical product approval, FDA can communicate available information to the public on the demographic profile of the study participants and on the demographic data subset analyses using a variety of mechanisms: initially with product labeling and publicly posted clinical reviews and later, once a product is on the market, with consumer updates, safety alerts, label changes, and other mechanisms, should this be necessary. 5 • Statutory differences in the regulatory framework for medical devices compared to those applicable to drugs and biologics account for differences in policies and practices across FDA centers with regard to submission and analysis of demographic data and public disclosure of information at the time a product is approved (e.g., timing of information, information release, and public documents). The information gathered for this report will become a starting point for developing an Action Plan, to be released next year, as required under Section 907. 6 FDA Report Collection, Analysis, and Availability of Demographic Subgroup Data for FDA-Approved Medical Products August 2013 Introduction For many decades, U.S. governmental initiatives have sought to identify the best ways to advance the three ethical principles that underlie the conduct of biomedical research: respect for persons, beneficence, and justice (which asks who ought to receive the benefits of research and bear its burdens). 3 Consistent with the principle of justice, FDA has a variety of statutory, regulatory, and policy-related tools that provide a framework for guiding medical product sponsors on the inclusion and analysis of demographic subgroups in clinical trials. This report describes these tools. However, scientific advances in understanding the specific genetic variables underlying disease and response to treatment are increasingly becoming the focus of modern medical product development as we move toward the ultimate goal of tailoring treatment to the individual or class of individuals. Thus, the broad, self-identified demographic subgroup categories used today may not adequately capture the complexity underlying responses to medical treatments. Nonetheless, it remains important that clinical trials include diverse populations, whenever possible and appropriate. On July 9, 2012, the President signed the Food and Drug Administration Safety and Innovation Act (FDASIA). 4 Section 907 of FDASIA, Reporting of Inclusion of Demographic Subgroups in Clinical Trial and Data Analysis in Applications for Drugs, Biological Products, and Devices called for a report with the following requirements: (1) In General Not later than 1 year after the date of enactment of this Act, the Secretary, acting through the Commissioner, shall publish on the Internet Web site of the Food and Drug Administration (FDA) a report, consistent with the regulations of the FDA pertaining to the protection of sponsors’ confidential commercial information as of the date of enactment of this Act, addressing the extent to which clinical trial participation and the inclusion of safety and 3 Ethical Principles and Guidelines for the Protection of Human Subjects of Research (The Belmont Report). Available at www.hhs.gov/ohrp/humansubjects/guidance/belmont.html. Accessed July 30, 2013. 4 Food and Drug Administration Safety and Innovation Act. Public Law No 112-144 (July 9, 2012). Available at http://www.gpo.gov/fdsys/pkg/PLAW-112publ144/pdf/PLAW-112publ144.pdf.. Accessed January 16, 2013. 7 effectiveness data by demographic subgroups including sex, age, race, and ethnicity, is included in applications submitted to the FDA, and shall provide such publication to Congress. (2) In General Not later than 1 year after the publication of the report described in subsection (a), the Secretary, acting through the Commissioner, shall publish an action plan on the Internet Web site of the Food and Drug Administration, and provide such publication to Congress. (2) Contents of the report – The report described in paragraph (1) shall contain the following: - A description of existing tools to ensure that data to support demographic analyses are submitted in applications for drugs, biological products, and devices, and that these analyses are conducted by applicants consistent with applicable Food and Drug Administration requirements and Guidance for Industry. The report shall address how the Food and Drug Administration makes available information about differences in safety and effectiveness of medical products according to demographic subgroups, such as sex, age, racial, and ethnic subgroups, to health care providers, researchers, and patients. - An analysis of the extent to which demographic data subset analyses on sex, age, race, and ethnicity is presented in applications for new drug applications for new molecular entities under section 505 of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 355), in biologics license applications under section 351 of the Public Health Service Act (42 U.S.C. 262), and in premarket approval applications under section 515 of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 360e) for products approved or licensed by the Food and Drug Administration, consistent with applicable requirements and Guidance for Industry, and consistent with the regulations of the Food and Drug Administration pertaining to the protection of sponsors’ confidential commercial information as of the date of enactment of this Act. - An analysis of the extent to which demographic subgroups, including sex, age, racial, and ethnic subgroups, are represented in clinical studies to support applications for approved or licensed new molecular entities, biological products, and devices. - An analysis of the extent to which a summary of product safety and effectiveness data by demographic subgroups including sex, age, race, and ethnicity is readily available to the public in a timely manner by means of the product labeling or the Food and Drug Administration’s Internet Web site. This report addresses demographic subgroup participation in clinical trials submitted to FDA in support of product applications; it describes demographic data collection, subset analysis, and public communication of this information. FDA’s next step will be to use the information gathered for this report as a starting point for developing an Action Plan also called for by Congress in Section 907. We look forward to input from patients, consumers, health care professionals, industry, and other stakeholders as we work to develop and implement the Action Plan. 8 Background To prepare this report, FDA convened an FDA-wide working group led by the Office of Women’s Health (OWH) in collaboration with the Office of Minority Health (OMH) and representatives from the Center for Biologics Evaluation and Research (CBER), the Center for Drug Evaluation and Research (CDER), the Center for Devices and Radiological Health (CDRH), the Office of Health and Constituent Affairs (OHCA), the Office of Pediatric Therapeutics (OPT), and other FDA offices. The working group undertook a systematic evaluation to complete the four tasks set out in the FDASIA legislation. 5 As described in Section 907 requirements, the assessment was to include information derived from sponsor applications, FDA reviews, and product labeling. The working group evaluated 72 applications approved during 2011 for new molecular entity 6 drug products, original biological products, and Class III devices (premarket approval). These are collectively referred to in this report as medical product applications unless otherwise specified. The applications evaluated for this report included 24 drugs, 11 biologics (5 approved by CBER and 6 approved by CDER − for 7 indications), 7 and 37 medical devices. Lists of the medical product indications approved in 2011 by CDER and CDRH and reviewed for this report are provided in Appendices 2 and 4, respectively; the indications for the five applications approved in CBER are listed in Part 1. The number of applications reviewed by the working group for this report generally reflects the average number of approvals per year. To provide some context, from fiscal year (FY) 2000 to FY 2010, CBER approved 64 biologics license applications, ranging from 2 to 10 per year. 8 For the 10-year period from calendar year 2002 through 2011, CDER averaged 24 novel new medicines, known as new molecular entities, approvals per year. 9 In 2011, CDER approved 30 NMEs (for 31 indications) representing the second highest total in the past 10 years (except 2004, during which CDER approved 36 new molecular entities). These new molecular entities approved by CDER included applications for both new drug applications (NDAs) and biologics license applications (BLA). For the 10-year period from calendar year 2002 through 2011, CDRH averaged 30 premarket approvals (PMA) per year. CDRH approved 37 PMAs in 2011, 5 Ibid. 6 A new molecular entity is an active ingredient that has never before been marketed in the United States in any form. Available at: http://www.fda.gov/drugs/informationondrugs/ucm079436.htm Accessed: May 15, 2013. 7 One product was approved for two different indications . 8 See Biological Approvals by Year: Available at: http://www.fda.gov/BiologicsBloodVaccines/DevelopmentApprovalProcess/BiologicalApprovalsbyYear/default.htm. Accessed: February 19, 2013. 9 See 2011 Novel New Drugs. Available at http://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/DrugInnovation/UCM293663.pdf. Accessed February 19, 2013. 9 representing the third highest total in the past 10 years (except 2004 and 2006, during which 47 and 39 PMAs were approved, respectively). 10 The findings in this report may not be generalizable to all FDA-approved medical products. However, selection bias was minimized by including all medical product applications approved during 2011, regardless of indication, trial size, and other factors (e.g., orphan designation) that could influence the findings. Nevertheless, because of the nature of medical product development, the 2011 approvals actually represent a culmination of various development activities that may span many years. Thus, the data analyzed for this report represent clinical trial data accrued over time. Practical factors also informed our decision to look at a one-year sample. The one-year statutory timeline to collect, analyze, write, clear, and publish this report was a challenge. For example, the working group had to manually abstract data from applications and other FDA documents to analyze the findings and compile the report within the allotted time frame. One approval decision for an application may have multiple sources of data that FDA staff needed to review to address the legislative requirements outlined in the section 907 of FDASIA. Report Organization Because the regulatory framework for drugs and biologics differs in important ways from that used for medical devices, the report is divided into two parts: Part 1 of the report discusses drugs regulated by CDER 11 and biologics regulated by CBER 12 and CDER. These centers regulate drugs and biologics according to the Federal Food, Drug, and Cosmetic Act (FD&C Act) 13 and the Public Health Service (PHS) Act, 14 respectively. Part 2 discusses medical devices, 15 which are regulated primarily by CDRH. 10 PMA Approvals. Available at http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/DeviceApprovalsandClearances/default.htm. Accessed 5/9/2013. 11 CDER regulates most therapeutic biologics products (e.g., recombinant products, such as enzymes and monoclonal antibodies) and drug products. 12 CBER regulates a variety of medical products including allergenics, blood and blood components and blood derivatives, certain medical devices, gene therapy, human tissues, cellular products, vaccines, and xenotransplantation products. 13 The FD&C Act defines drugs, in part, by their intended use, as "articles intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease" and as "articles (other than food) intended to affect the structure or any function of the body of man or other animals" (FD&C Act, sec. 201(g)(1). Available at http://www.gpo.gov/fdsys/pkg/USCODE-2010-title21/pdf/USCODE-2010-title21-chap9-subchapII-sec321.pdf. Accessed April 4, 2013. 14 The Public Health Service Act (PHS Act) 1944. Available at http://www.fda.gov/regulatoryinformation/legislation/ucm148717.htm. Accessed March 7, 2013. 15 Medical devices range from simple tongue depressors and bedpans to complex programmable pacemakers with micro-chip technology and laser surgical devices. Additionally, medical devices include in vitro diagnostic products, such as general purpose lab equipment, reagents, and test kits, which may include monoclonal antibody technology. Certain electronic radiation-emitting products with medical application and claims meet the definition of medical device. Examples are diagnostic ultrasound products, x-ray machines, and medical lasers. 10 Following a brief overview of the tools FDA uses to facilitate the collection and analysis of demographic subgroup data, this report discusses the findings of the working group. Tools to Ensure Analysis of Demographic Information FDA uses a variety of tools to ensure the collection, submission, and analysis of demographic data with clinical trial data. Although there is no statutory or regulatory requirement to include demographic subgroups as participants in clinical trials, FDA guidance documents encourage, 16, 17,18 and regulations clearly require, presentation and inclusion of analyses of demographic data in marketing applications (see Appendix 1, Table 1). In 1985, FDA issued regulations on the content and format of new drug applications (21 CFR 314.50 19) that require the presentation of effectiveness data by gender, 20 age, and racial subgroups and the identification of dosage modifications for specific subgroups. Those same regulations require safety data to be presented by gender, age, and racial subgroups, and when appropriate, safety data from other subgroups of the population of patients treated must also be presented. In 1998, FDA amended its investigational new drug application regulations (21 CFR 312.33 21) to require that data on the participation in clinical trials be presented in annual reports by age group, gender, and race (see additional regulations in Appendix 1, Table 1). FDA guidance documents provide additional policy recommendations relevant to the collection and analysis of relevant subgroup data in clinical trials (see the list in Appendix 1, Table 2). For example, in 1993, FDA issued guidance to encourage inclusion of women of childbearing potential in phase 1 and early phase 2 trials; made recommendations for the assessment of potential pharmacokinetic (PK) differences between genders; and encouraged sponsors to collect sex-related data during research and development and analyze the data for sex effects. 22 In 2011, FDA issued draft guidance on evaluating sex differences in medical device studies, outlining 16 FDA guidance E7 Studies in Support of Special Populations: Geriatrics, Questions and Answers. Note: FDA guidance documents are available at http://www.fda.gov/ForIndustry/FDABasicsforIndustry/ucm234622.htm 17 FDA guidance Guideline for the Study and Evaluation of Gender Differences in the Clinical Evaluation of Drugs 18 There is evidence that guidance has been successful in getting good sex demographic subgroup representation, as noted in Appendix 1 of FDA’s Guideline for the Study and Evaluation of Gender Differences in the Clinical Evaluation of Drugs. One example is a follow-up GAO study from 2000, in which GAO noted that women are now included in clinical research at rates proportional to their representation in the population. (United States General Accounting Office, Report to Congressional Requesters. Women’s Health, NIH Has Increased Its Efforts to Include Women in Research). Available at http://www.gao.gov/archive/2000/he00096.pdf. Accessed April 12, 2013. 19 Content and format of an application, 21 CFR Sect. 314.50 (2012). Available at http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=314.50. Accessed May 9, 2013 20 The terms sex and gender have been used interchangeably in some FDA documents. 21 Annual reports, 21 Code of Federal Regulations (CFR) Sect. 312.33 (2012). Available at http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=312.33. Accessed May 9, 2013. 22 FDA guidance Guideline for the Study and Evaluation of Gender Differences in the Clinical Evaluation of Drugs. 11 FDA’s expectations on sex-specific patient enrollment, data analysis, and reporting of study information. 23 The guidance recommends that data from such studies be appropriately analyzed for sex differences. Once marketing applications have been submitted to FDA for review, the Agency uses a variety of tools as part of the product review process to make sure sponsors have met relevant requirements. Examples of FDA reviewer tools include reviewer checklists, review templates, and a variety of internal standard operating procedures (see Appendix 1, Table 3). 24 23 FDA draft Guidance for Industry and Food and Drug Administration Staff, Evaluation of Sex Differences in Medical Device Clinical Studies. 24 Reviewer checklists ensure that all relevant sections and information, such as demographic information of participants, are included in an application prior to review. The review templates are structured outlines with an annotated table of contents used in preparing a marketing application review. Review templates aid FDA reviewers in organizing review content for consistency in documentation, using good review practices. FDA standard operating procedures provide reviewers with general operating standards for reviewing product applications and their respective supplements, product labeling, and annual reports. These operating procedures are directed toward FDA staff in the performance of their daily activities. 12 Part 1: Drugs and Biologics A. Introduction This section reports on the following with regard to drugs and biologics: • The extent to which these applications include demographic data subset analyses on age, sex, and race/ethnicity population categories (subset analysis) The extent to which demographic subgroups are represented in clinical trials • The public availability of a summary of product safety and effectiveness data by demographic subgroups and the timeliness of availability This section discusses new molecular entity drugs and new biologic licensing applications approved in the Center for Biologics Evaluation and Research (CBER) and the Center for Drug Evaluation and Research (CDER) during 2011. During this time period, 35 new drugs and biologics were approved or licensed for 36 new indications (indication is the approved use). 25 A list of the 31 indications (24 drug indications and 7 biologics indications) approved in CDER that were reviewed for this report is provided in Appendix 2. The five biologics applications approved in CBER that were reviewed for this report were for the following indications: • For use in military populations 17 through 50 years of age for active immunization for the prevention of febrile acute respiratory disease caused by Adenovirus Type 4 and Type 7 (prophylactic vaccine) • For the routine prophylactic treatment of Congenital Factor XIII deficiency (orphan designation) • For the treatment of clinical signs of scorpion envenomation (orphan designation) • For improvement of the appearance of moderate to severe nasolabial fold wrinkles in adults 26 • For the detection of delayed-type hypersensitivity to C. immitis in individuals 18–64 years of age with a history of pulmonary coccidioidomycosis (orphan designation) The working group evaluated marketing applications submitted to FDA, clinical review documents, approved product labeling (package inserts), and other documents on FDA’s web site to determine the extent to which demographic information was collected, analyzed, and included 25 One product was approved for 2 different indications. 26 This is an autologous product (i.e., derived using the patient’s own cells). 13 in the sponsors' marketing applications to FDA and made publicly available once an application was approved. The demographic categories used in Part 1 include age, sex, race, and ethnicity, defined as: Age: Pediatrics (age ≤16 years), geriatrics (age ≥65 years) 27 Sex: Male, female Race: White, Black or African-American, Asian, American Indian or Alaska Native,* Native Hawaiian or Other Pacific Islander.* *Since the percentages of these categories were low for each of the drug and biologics applications, they were summarized together in this report. Ethnicity: Defined as Hispanic or Latino or not. Note on Ethnicity: Although ethnicity was addressed in many of the documents, for drugs and biologics approved in CDER, it was not analyzed for this report. This is because some applications reported race and ethnicity as one item, rather than separately (no requirement exists for sponsors to collect these data separately). 28 The largest ethnic group defined was Hispanic. However, in some applications, the population was further defined as subpopulations and in others, it was not. Absence of ethnic distinctions in an application does not necessarily mean that Hispanic/Latino or other ethnic groups of patients were not studied; rather, these patients were categorized by their race, not their ethnicity. The Office of Management and Budget (OMB) acknowledges that the categories in this classification are social-political constructs and should not be interpreted as being scientific or anthropological in nature. 29 B. Demographic Subgroup Information Contained in Applications The following discussion presents the findings in more detail, with accompanying graphic representations of the demographic data reported and analyzed, first for applications approved in CDER, then for applications approved in CBER. 27 According to 21 CFR 201.57(f)(9)(i), the pediatric age group is defined as "birth to 16 years, including age groups often called neonates, infants, children, and adolescents." FDA states in guidance, however, that the Best Pharmaceuticals for Children Act (BPCA) defines pediatric studies to include studies in “all pediatric age groups including neonates in appropriate cases, in which a drug is anticipated to be used.” For the purposes of satisfying the requirements of the Pediatric Research Equity Act (PREA), the appropriate age ranges to be studied may vary, depending on the pharmacology of the medical product, the manifestations for the disease in various age groups, and the ability to measure the response to therapy. In general, however, the pediatric population includes patients aged “birth to 16 years, including age groups often called neonates, infants, children, and adolescents.” 28 Sponsors interested in collecting ethnicity and race separately can follow the recommendations for doing so in FDA guidance (i.e., Collection of Race and Ethnicity Data in Clinical Trials). 29 Office of Management and Budget, Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity. Available at www.Whitehouse.gov/omb/fedreg_1997standards/ Accessed March 12, 2013. 14 Findings for Applications Approved in CDER For the drug and biologics applications approved in CDER in 2011, the working group collected data from the key clinical studies, identified from approved product labeling as those studies that provide evidence of efficacy, safety, and benefit–risk to support FDA’s approval decisions. These studies, sometimes referred to as pivotal trials, are clinical investigations designed to collect definitive evidence of the safety and effectiveness of a medical product for a specified intended use, typically in a statistically justified number of patients. The working group found that all of the applications approved in CDER provided information on the demographic composition of the clinical trials. The applications also addressed demographic subset analyses in the summaries of safety and/or efficacy, except for one small study of a product to treat a rare disease. For this one exception, FDA determined that the analysis for a study of fewer than 60 patients with a rare disease was unnecessary, given that the number of patients was too small for subgroup analysis. For the remaining applications, when the number of patients in a demographic category was sufficient for subset analysis, data were presented in the application. Detailed information with graphs of this analysis is found in Section C below. 1. CDER – Age Composition All of the applications approved in CDER and evaluated by the working group reported age composition (geriatrics and pediatrics). • Geriatrics Overall, the findings showed that the percentage of geriatric patients participating in clinical studies varied by indication and tended to reflect the prevalence of the disease in the geriatric population. For example, acute lymphoblastic leukemia (ALL), Hodgkin’s lymphoma, LennoxGastaut syndrome (a form of childhood-onset seizure disorder), and systemic lupus erythematosus (SLE) occur more commonly in younger patients; thus, geriatric representation would be expected to be low in these clinical studies. Conversely, age-related macular degeneration (AMD) is a disease that occurs predominantly in older people, and the percentage of geriatric patients participating in AMD clinical studies would be expected to be high. Figure 1-1 shows the results for clinical trial demographic composition by geriatric age. In this figure, the percentage of trial participants who were 65 years or older is plotted according to the product indication. This approach was taken because product indication is often highly relevant to the composition of patients participating in the clinical studies. 15 Figure 1-1: Trial Composition – Geriatrics Figure 1-1 Abbreviations: COPD = chronic obstructive pulmonary disease; DVT = deep venous thrombosis; HTN = hypertension; TE/ACS = thrombotic events in acute coronary syndrome; T2DM = type 2 diabetes mellitus; ALCL=anaplastic large cell lymphoma; ALL = acute lymphoblastic leukemia; Hodgkin = Hodgkin’s lymphoma; NSCLC = anaplastic lymphoma kinase-positive non-small cell lung cancer; Prostate CA = prostate cancer; Thyroid CA = medullary thyroid cancer; C. Diff = Clostridium difficile-associated diarrhea; Hep C = hepatitis C; SLE = systemic lupus erythematosus; MRI = intravenous use in diagnostic magnetic resonance imaging; RLS = restless leg syndrome; AMD = age-related macular degeneration Note: The geriatric data graph excludes eight drugs and biologics − even though the documents for them did include age information − because the application summaries provided age information as a median with range (e.g., median age 60 with a range of 40 to 87 years) or used an age grouping different from ≥65 years (e.g., ≥60 years), and the percentage of patients >65 years could not be found. • Pediatrics Pediatric demographic composition graphs were not generated because only three of the drugs and biologics were studied in pediatric patients. The small number of pediatric patients studied is an expected finding because half of the drugs and biologics have an indication that rarely or never occurs in children (e.g., age-related macular degeneration). In addition, new drugs and biologics, 16 like those reviewed for this report, are generally studied first in adults before exposing children to unknown risks. The Pediatric Research Equity Act (PREA) 30 may require additional study in children in the postmarket period. For 12 of the applications reviewed for this report, additional postmarket studies in children were identified in the product approval letters as studies to be submitted as supplements to the original application (approval letters are publicly available on Drugs@FDA). 31 2. CDER – Sex Composition All of the applications approved in CDER during 2011 and examined for this report reported trial composition by sex. Overall, and similar to the findings by age, the percentage of patients by sex who participated in clinical studies tended to reflect the prevalence of the disease in men and women. For example, there was no female representation in the prostate cancer trial, whereas SLE, which is predominantly a disorder of women, had a high percentage of female participants. Figure 1-2 shows the results for clinical trial demographic composition by female sex, plotted according to the drug or biologics’ indication. 30 Pediatric Research Equity Act of 2003. Public Law No 108-155 (December 3, 2003). Available at http://www.gpo.gov/fdsys/pkg/PLAW-108publ155/html/PLAW-108publ155.htm. Accessed November 29, 2012. 31 Drugs@FDA. Available at http://www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm. Accessed November 28, 2012. 17 Figure 1-2: Trial Composition – Sex Figure 1-2 Abbreviations: COPD = chronic obstructive pulmonary disease; DVT = deep venous thrombosis; HTN = hypertension; TE/ACS = thrombotic events in acute coronary syndrome; T2DM = type 2 diabetes mellitus; ALCL=anaplastic large cell lymphoma; ALL = acute lymphoblastic leukemia; Hodgkin = Hodgkin’s lymphoma; iron overload = transfusional iron overload; NSCLC = anaplastic lymphoma kinase-positive non-small cell lung cancer; Prostate CA = prostate cancer; Thyroid CA = medullary thyroid cancer; C. Diff = Clostridium difficile-associated diarrhea; Hep C = hepatitis C; HIV = human immunodeficiency virus; HAE = hereditary angioedema; MDD = major depressive disorder; SLE = systemic lupus erythematosus; MRI = intravenous use in diagnostic magnetic resonance imaging; RLS = restless leg syndrome; SPECT = single photon emission computed tomography imaging for Parkinsonian syndromes; AMD = age-related macular degeneration 3. CDER – Race Composition All of the applications approved in CDER and examined for this analysis reported trial composition by race. The overall findings showed that Whites represented a high percentage of trial participants, which may in part reflect lower percentages of non-White racial demographic subgroups in the U.S. population (see Appendix 3). For some of the indications, race composition in the trials was consistent with the disease prevalence (e.g., melanoma and head lice are more prevalent in Whites). However, for other indications, race composition was not consistent with disease prevalence in the U.S. population, and African American representation was low relative to the African American population with the disease. For example, for type 2 diabetes mellitus (T2DM), African American representation in clinical studies was less than 5%, even though African Americans make up approximately 13% of the U.S. population and have a higher prevalence of T2DM. 32,33 32 See Appendix 3 for U.S. Census Bureau 2011 statistics on People Quick Facts for the percentage of U.S. population by race. 18 Table 1-3 shows the results for clinical trial demographic composition by race presented by the percentage of trial participants who were White, African-American, Asian or Other, according to the drug or biologic indication. 33 U.S. Department of Health and Human Services, Office of Minority Health. Diabetes and African Americans.. Available at http://minorityhealth.hhs.gov/templates/content.aspx?ID=3017. Accessed April 4, 2013. 19 Table 1-3: Trial Composition - Race Indication White Asian Other 84% Black/African American 2% COPD (1) 11% 3% COPD (2) 87% 2% 7% 4% DVT 86% 1% 7% 5% HTN 70% 19% 2% 9% TE/ACS 92% 1% 6% 1% T2DM 67% 2% 31% 0% ALCL 83% 12% 2% 3% ALL 78% 10% 2% 10% Hodgkin 87% 5% 7% 1% Iron Overload 73% 1% 17% 8% Melanoma (1) 99% 1% 0% 0% Melanoma (2) 100% 0% 0% 0% Myelofibrosis 88% 2% 2% 8% NSCLC 63% 3% 30% 4% Prostate CA 93% 4% 2% 2% Thyroid CA 95% 1% 3% 2% C.Diff 90% 9% 1% 0% Head Lice 92% 0% 2% 6% HepC (1) 83% 14% 2% 2% HepC (2) 87% 9% 2% 3% HIV 61% 23% 13% 3% HAE 94% 0% 0% 6% Organ Rejection 67% 11% 8% 15% SLE 52% 12% 17% 19% Lennox-Gastaut 67% 11% 20% 2% MDD 81% 14% 2% 3% MRI 46% 5% 45% 3% RLS 95% 2% 1% 2% Seizures 92% 3% 1% 4% SPECT 99% 1% 0% 0% AMD 85% 0% 11% 4% Table 1-3 Abbreviations: COPD = chronic obstructive pulmonary disease; DVT = deep venous thrombosis; HTN = hypertension; TE/ACS = thrombotic events in acute coronary syndrome; T2DM = type 2 diabetes mellitus; ALCL=anaplastic large cell lymphoma; ALL = acute lymphoblastic leukemia; Hodgkin = Hodgkin’s lymphoma; iron overload = transfusional iron overload; NSCLC = anaplastic lymphoma kinase-positive non-small cell lung cancer; Prostate CA = prostate cancer; Thyroid CA = medullary thyroid cancer; C. Diff = Clostridium difficile-associated diarrhea; Hep C = hepatitis C; HIV = human immunodeficiency virus; HAE = hereditary angioedema; SLE = systemic lupus erythematosus; MDD = major depressive disorder; MRI = intravenous use in diagnostic magnetic resonance imaging; RLS = restless leg syndrome; SPECT = single photon emission computed tomography imaging for Parkinsonian syndromes; AMD = age-related macular degeneration 20 Findings for Applications Approved in CBER For the five biologics applications reviewed, the working group collected data from the approved product labeling and the clinical reviews. The working group found that all of the applications approved in CBER provided information on the demographic composition of the clinical trials, including the number of subjects, age, sex, and race for the pivotal studies to support approval of their products. When viewing the graphic representations that follow, it is important to note that for orphan products, the sample sizes are relatively small 34 so representations of different subgroups on the graph may appear large but, in fact, may only differ by one or two patients. 4. CBER – Age Composition In all five applications, the ranges of the participant ages were reported. Pediatric patients were enrolled in studies in three of the applications, and geriatric patients were enrolled in studies in two. Figure 1-4 shows the age range by indication for the applications approved in CBER. Figure 1-4: Age Range by Submission (CBER)-Efficacy Trial Composition 34 The Orphan Drug Act defines a rare disease as a disease or condition that affects fewer than 200,000 persons in the United States. Orphan Drug Act. Public Law No 97-414. Available at http://www.fda.gov/RegulatoryInformation/Legislation/FederalFoodDrugandCosmeticActFDCAct/SignificantAmend mentstotheFDCAct/OrphanDrugAct/default.htm Accessed April 4, 2013. 21 Two applications subject to PREA received full pediatric study waivers, so studies in the pediatric subpopulations are not required. Full waivers are granted in many cases when the disease or condition is not likely to occur in the pediatric population. For example, for the prophylactic vaccine, which is indicated for the military population, it is expected that pediatric and geriatric subgroups would not have been enrolled. This is reflected in the product labeling, and the age ranges are stated in the indication and labeling where appropriate. The remaining three applications were not subject to PREA due to orphan designation; however, two included pediatric patients in the study cohort, as noted in Figure 1-4, since the disease occurs in the pediatric population. When significant safety concerns or insufficient data exist for certain subpopulations, this information is made available in the appropriate section (i.e., 8.4 Pediatric Use or 8.5 Geriatric Use) of the product labeling. For example, for the biologic indicated for the improved appearance of nasolabial fold wrinkles, the label states under the heading, 8.5 Geriatric Use: Clinical studies … did not include sufficient numbers of subjects aged 65 years and older to determine whether they respond differently from younger subjects. [See Clinical Studies (14)]. Additionally, when outcomes are analyzed by age, these are stated in the label. The following language is present in two different product labels, both of which are orphan products: "There were no apparent differences in the safety profile in children as compared to adults" and “The efficacy and safety … is comparable in pediatrics and adult patients." 5. CBER – Sex Composition All five applications approved in CBER described trial populations by sex (Figure 1-5). For the nasolabial fold wrinkle indication, more than 90% of the studied population was female. In the remainder of the studies, the percentage of participants was generally evenly distributed between male and female participants. 22 23 Adenovirue Types 4 and Neeolebiel fold wrinkles Scorpion evenornetion Indication Coeeidioidee immitie Congenital FaetorXIll Deficiency Pereent Female -1 OJ U1 *4 03 100? Figure 1-5 Composition by Submission (CBER)-Efficacy Trial Composition 6. CBER – Race Composition All five applications approved in CBER during 2011 provided racial and/or ethnicity data. Although a large percentage of the enrolled patients were White, this was not always the case. White participants comprised the majority of all racial subgroups enrolled for four of the five applications. For one orphan product, African American or Black patients made up the highest percentage among the racial subgroups enrolled. In the application for a nasolabial fold wrinkle indication, the percentage of Whites exceeded 90%, and the labeling reflects the race composition. The labeling under the heading, 8.6 Race, states: Clinical studies … did not include sufficient numbers of subjects in non-White populations to determine whether they respond differently from the population studied. The racial and ethnic study population enrollment can also be influenced by geographic distribution of the disease (e.g., scorpion envenomation). The population enrollment also can be influenced by the indication and target population. For example, in the product for prevention of respiratory disease caused by adenovirus types 4 and 7 for the military population, approximately 18%, or almost one fifth of the subjects enrolled were African American, which is representative of the U.S. military population 35 (see Figure 1-6). 35 United States Department of Defense. Remembering the Legacy, African Americans in the Military. Available at http://www.defense.gov/home/features/2007/blackhistorymonth/index.html Accessed April 4, 2013. 24 Figure 1-6: Racial?Ethnic Composition by Submission (CBER)-Efficacy Trial Composition 100 I White 90 - African American I Black - Asian 80 Cl Hispanic - Other Indication 25 C. Subset Analyses in Applications and Public Availability of Data Findings for Applications Approved in CDER Drug and biologics applications are reviewed by CDER scientists and their reviews are publicly available on FDA’s Web site (i.e., posted on Drugs@FDA) if the new drug or biologic is approved for marketing. Labeling is likewise posted for approved drugs and biologics. FDA approved labeling refers to the official description of a drug or biologic and includes information on how the drug or biologic should be prescribed (e.g., dose, indication, patient population), side effects and safety information, and instructions for use. The labeling also includes summary results of the key clinical studies (evidence of effectiveness) that supported drug or biologic approval. The working group examined labeling posted on Drugs@FDA for demographic subset analyses of effectiveness and safety, including clinical pharmacology (clinical pharmacology addresses how drugs and biologics are metabolized by the body, drug interactions, and other information). If the information was not found in the labeling, the FDA medical officer, biostatistician, and/or clinical pharmacologist reviews were consulted. The rationale for this approach was that all information in labeling underwent review (and would, therefore, appear in FDA reviews), but reviews would also contain additional information that was not included in labeling. This is because FDA-approved product labeling is designed to communicate information important to the safe and effective use of the product, not to provide a comprehensive catalogue of all available information. For each demographic category (age, sex, and race), we asked two questions for this report about each subset analysis (safety, efficacy, clinical pharmacology): 1. Does a public document address demographics? This question was answered “no” if no mention of the topic was found (e.g., no mention of subset analysis of safety by age). 2. Does a public document contain data about subset analysis for each of the demographic categories? This question was answered ‘no’ if the document stated that representation by a demographic category was too low for meaningful analysis, or similar explanation (e.g., if a small trial had only a few percent non-White participants, the answer to this question was “no data”). Publicly available documents on Drugs@FDA (labeling and/or reviews) provided demographic information for all of the CDER-approved drugs and biologics examined for this report. Information included the demographic composition of trials, and, when sufficient demographic subset numbers were enrolled, subset information was provided. The following sections and figures show the subset information for each of the demographic categories (age, sex, and race), including whether the information was found in labeling, one or more reviews, or both. 26 1. CDER−Public Availability of Age Subset Data Sections on geriatric and pediatric age subset information are required in product labeling 36; thus, all of the applications reviewed for this report addressed age subsets in labeling. • Geriatrics For all of the applications, the age composition of the trials was reported in publicly available documents, and all labeling had required geriatric labeling subsections. Figure 1-7 shows the results for subset analysis by the geriatric age subset. This figure shows whether geriatrics was addressed with data or without data for each type of subset analysis, including clinical pharmacology, efficacy, and safety analyses, as follows: • • • • The first bar in each graph shows the number of drugs and biologics with publicly available information in labeling and at least one FDA review. The second bar shows the number of applications that had information in an FDA review, but not in labeling. The last bar shows the number (if any) for which no publicly available information was found on Drugs@FDA. The sum of all the bars in each graph equals the total number of approved drug and biologic indications in this report (31). The blue part of each bar indicates the presence of data; the red part indicates that the topic was addressed, but without data (e.g., the Geriatrics labeling subsection would state that few patients ≥65 years old were studied). The first graph shows that geriatric clinical pharmacology information was found in labeling and in at least one review for 24 of the 31 indications. For an additional six indications, information was not in labeling, but was found in at least one review. No publicly available information was found for the one remaining application. 36 Requirements on content and format of labeling for human prescription drug and biological products. 21 CFR Section 201.56. Available at http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=201.56. Accessed May 9, 2013. 27 Figure 1-7: Public Availability of Geriatric Subset Analysis Note: There were 31 indications for 30 drugs and biologics. In these figures, drugs is used to refer to NME drugs and biologics approved or licensed by CDER. When the total number of indications = 31, the graph range begins at 1 rather than 0; on those graphs, 1 appears as the straight line, rather than a bar above the line. • Pediatrics For all of the applications, the age composition of the trials was available in publicly available documents, and all labeling had required pediatric labeling subsections, but most did not contain pediatric clinical data. As was noted earlier, this was an expected finding. 28 2. CDER−Public Availability of Sex Subset Analysis For all of the reviewed applications, the sex composition of the key clinical studies was available in publicly available documents and almost all had sex subset analysis data in publicly available documents. Figure 1-8 shows the results for subset analysis by sex (male or female). This figure is organized exactly as described for the age subset analysis figure, except that the total number of drugs and biologics equals 30 instead of 31 because one product was indicated exclusively for males for the treatment of prostate cancer, and no analysis by sex would have been performed. Figure 1-8: Public Availability of Sex Subset Analysis Note: There were 30 indications for 29 drugs and biologics. In the figures, the term drugs refers to both NME drugs and biologics approved in CDER. One sex-specific indication was excluded from the analysis. These graphs show that for most of the reviewed applications, clinical pharmacology sex subset data were included in labeling (24 of 30), but not efficacy and safety subset analyses. However, most had efficacy and safety information in the reviews (thus, for the efficacy and safety graphs, 29 the second bar, review only, is larger than the first bar, review and labeling). In the efficacy graph, adding the two bars (6 + 23) shows that for almost all (29 of 30), sex subset data were in at least one publicly available document. We expected fewer labels to include sex demographic information compared to pediatrics and geriatrics, which must be addressed in labeling regardless of whether there is informative data (sex demographic labeling is not a required subsection). Thus, inclusion of sex demographic information in labeling is decided on during the review process based on whether the available information would be informative to prescribers. To a large extent, whether it is included depends on whether there are sufficient numbers of males and females for meaningful subset analysis or whether relevant differences based on sex were noted in clinical studies. 3. CDER−Public Availability of Race Subset Analysis For all of the reviewed applications, race composition of the trials was made publicly available in documents. Race subset analysis data were publicly available for about 50% (safety analysis) and 80% (efficacy and clinical pharmacology) of the products. Figure 1-9 shows the results for subset analysis by race. This figure is organized as described for the age subset analysis figure. 30 Figure 1-9: Public Availability of Race Subset Analysis Note: There were 31 indications for 30 drugs and biologics. In the figures, the term drugs refers to both NME drugs and biologics approved in CDER. When the total number of indications=31, the graph begins at 1 rather than 0; on those graphs, 1 appears as the straight line rather than a bar above the line. These graphs show that approximately half of the applications included clinical pharmacology race subset data in labeling (18 of 31), but few had efficacy and safety subset analyses in the labeling. However, most applications had efficacy and safety subset data in the reviews. For example, the efficacy graph shows that most (25 of 31) had efficacy subset data in at least one publicly available document (7 in labeling + 18 in reviews posted on FDA’s Web site). Safety subset data were present for 15 of the 31 indications. As with sex demographic subset information, it is expected that fewer labels will include race demographic information compared to pediatrics and geriatrics because these are not required sections for labeling. Thus, inclusion of race demographic information in labeling is decided on during the review process, based on whether the available information would be informative to prescribers. To a large extent, this depends on whether sufficient numbers of non-White patients were available for meaningful subset analysis. 31 An absence of racial subset data is expected for drugs and biologics approved to treat diseases that primarily affect Whites (e.g., head lice infestation is less common in African Americans, as is melanoma). For many of the applications, there were too few African-American or Black patients in the trials to enable meaningful subset analysis (e.g., see prostate cancer in Figure 3). It is notable that the approval letters for 4 of the 31 indications included a postmarket commitment or requirement related to race: transfusional iron overload (Iron Overload), single photon emission computed tomography imaging for Parkinsonian syndromes (SPECT), SLE, and hepatitis C. Overall, the findings for the applications approved in CDER and reviewed for this report show that, as expected, FDA reviews address demographics more often than does labeling. This is because labeling is not a comprehensive summation of all available information. Reviews address demographics even when these categories have no effect on safety and/or efficacy, or there is insufficient data to make that determination. Because FDA reviews are posted on the Internet after a new drug or biologic is approved, the public has access to all of this information. The most efficient way to access this review information is to go to Drugs@FDA, type in the product name, click on Reviews, and then choose the review of choice. For example, if one is interested in demographic differences in metabolism and dosing of a drug or biologic, look at the Clinical Pharmacology review; for demographic subset information, consult the Medical Officer/Clinical reviews or the Biostatistician reviews. These reviews have tables of contents to aid in efficient information access. Ultimately, however, clinically important information is most readily available in the product labeling, which will include a patient-centered Medication Guide, if required, for dissemination of important safe use information to patients. These are accessible at the same site by clicking Medication Guide instead of Reviews. As this section shows, in some cases, there was no subset analysis of safety and effectiveness data available publicly, either in FDA reviews or product labeling. There were a number of reasons for this. When a product was intended to treat a disease occurring primarily in one subgroup, there were too few patients from other demographic categories for meaningful inclusion in the analysis (e.g., few male patients in a study for a disease that rarely affects men). This also occurred for the rare disease products when the totality of all affected patients precluded subset analysis because the key clinical studies were, by necessity, small (one third of the applications examined for this report were for rare disease indications). 32 Findings for Applications Approved in CBER Descriptions of clinical trial populations by age, sex, and race were included in the labeling and/or clinical reviews of the five products, and the information is publicly available. For products labeled for pediatric use, this information is posted according to the pediatric tracking requirements under Food and Drug Administration Amendments Act of 2007 (FDAAA). 37 The issue of subset analyses was addressed, but in some applications not all subsets were analyzed; the reasons varied. For three of the five applications (product for adenovirus Types 4 and 7, detection of C. immitis, and scorpion envenomation treatment), when the overall effectiveness reflected an overwhelmingly positive response (e.g., greater than 95% effectiveness), the high rate of effectiveness obviated the necessity of additional subgroup analysis. In one application (nasolabial fold wrinkles), the overwhelming number of subjects were White and female (>90%) and thus subset analysis of other subgroups would not have been meaningful. In one application (for Congenital Factor XIII) for an orphan product, the number of subjects studied was too small for a subgroup analysis to be meaningful (N=13). FDA regulations 38 related to biologics do not require sponsors of new biologics applications to present a summary of safety and effectiveness data by demographic subgroups (age, sex, race), or an analysis of whether modifications of dose or dosage intervals are needed for specific subgroups. Nonetheless, FDA provides recommendations to sponsors through its FDA guidance documents to submit such data (see Appendix 1, Table 2). 37 Food and Drug Administration Amendments Act of 2007. Public Law No.110-85 (September 27, 2007). Available at http://www.fda.gov/RegulatoryInformation/Legislation/FederalFoodDrugandCosmeticActFDCAct/SignificantAmend mentstotheFDCAct/FoodandDrugAdministrationAmendmentsActof2007/FullTextofFDAAALaw/default.htm. Accessed April 4, 2013. 38 Biological Products, 21 CFR Part 600. Available at http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/cfrsearch.cfm?cfrpart=600. Accessed May 9, 2013. 33 D. Summary of Findings for Part 1: Drugs and Biologics In summary, the findings based on the review of marketing applications for drugs and biologics submitted to FDA and reviewed for this report were as follows: 39 • Sponsors collected and analyzed demographic trial composition data and submitted them to FDA in their applications. • FDA reviewers addressed demographics in their reviews, which are available on FDA’s Web site. • Clinically meaningful demographic subset analysis information was included in approved labeling, which is publicly available at the time the drug or biologic is marketed in the United States. • The extent to which demographic subgroups were represented in the clinical trials showed that, in general, participation by age and sex tended to reflect the disease indication studied. • Whites represented a high percentage of trial participants. In many cases, other racial subgroups were underrepresented. Ethnicity data varied in terms of whether ethnicity was captured separately or in combination with race categories, and, although provided consistent with current guidance, 39 the data could not be analyzed in a meaningful way for this report. FDA guidance for industry Collection of Race and Ethnicity Data in Clinical Trials for FDA Regulated Products. 34 Part 2: Medical Devices A. Introduction This section addresses the requirement to analyze and report on the following with regard to medical devices: • The extent to which these applications include demographic data subset analyses on age, sex, and race/ethnicity population categories (subset analysis) • The extent to which demographic subgroups are represented in clinical trials • The public availability of a summary of product safety and effectiveness data by demographic subgroups and the timeliness of availability The working group evaluated original premarket approval (PMA) applications for medical devices (including diagnostic devices) approved in 2011 in FDA’s CDRH. PMA supplement applications used to obtain approval for changes or modifications to an approved device were not included in this assessment. 40 A total of 37 PMA applications were approved in 2011. These included 46 pivotal clinical studies to support the applications. Six PMAs relied on data used in other PMAs (four PMAs relied on one dataset and two PMAs on another). Thus, 33 PMAs contained unique study datasets. The working group counted study populations used by more than one PMA only once for the description of the demographics of study participants. All 37 PMAs were reviewed for the questions about demographic data subset analyses and the communication of these results. A list of the applications for medical devices approved in CDRH that were reviewed for this report is provided in Appendix 4, with their specific indications for use. The working group evaluated PMA applications, approved product labeling, other documents on FDA’s Web site, and other relevant review documents to determine the extent to which sponsors collected, analyzed, and submitted to FDA demographic information. In addition, the working group determined the extent to which demographic information was made publicly available. The demographic categories discussed in this report include age, sex, ethnicity, and race. 40 PMA supplements are submitted for review and approval by FDA when there may be a change affecting safety or effectiveness of an already approved device, such as to include a new indication for device use or labeling changes. Additional demographic data from clinical studies may be submitted to FDA in support of these supplemental applications. However, this was beyond the scope of this report because the results are not available at the time of the original marketing application submission. 35 CDRH uses the following categories for each demographic subgroup: Age: Sex: Race: Ethnicity: Mean Age, Standard Deviation, Median Age, Minimum Age, Maximum Age (pediatric age groups for devices are defined up to, but not including, the twentysecond birthday) 41 Male, Female, Not Stated White, Black or African American, Asian, American Indian or Alaskan Native, Native Hawaiian or Other Pacific Islander, Other, Not Stated Hispanic or Latino, or not The working group reviewed Summaries of Safety and Effectiveness Data (SSEDs), 42 device labeling, review documentation, and the product applications to extract demographic information. Similar to the approach used for CDER and CBER, CDRH data were collected initially from the publicly available labeling and SSED. When demographic information was not found in public documents, the working group turned to non-public review documentation and the product applications. Trial composition demographics were recorded for investigational and control groups and for total study participants (the sum of participants in the investigational and control groups). For studies that used a non-active control group (e.g., historical or literature control group), wherein control participants were not evaluated during the same timeframe as the investigational group or in the same study, trial composition demographics were recorded for the investigational device group only. Demographic subset analyses were recorded if there was: (1) a clear analysis of primary and secondary endpoints comparing the relevant subpopulation (sex, age, ethnicity or race); (2) a presentation of results by sex, age, ethnic, or race subpopulations; or (3) a statement that a subset analysis was done that summarizes the results. Public documents describe subgroups that were or were not represented in the clinical trial to support device approval. A data subset analysis was marked as Not Applicable if the device had a sex-, age-, ethnicity-, or race-specific indication. Of the 37 PMA applications, 5 PMAs were approved with sex-specific indications, and 7 were approved with age-specific indications; no applications were approved for race- or ethnicityspecific indications. 41 FDA Guidance for Industry and FDA Staff on Premarket Assessment of Pediatric Medical Devices. Note: FDA guidance documents are available at http://www.fda.gov/ForIndustry/FDABasicsforIndustry/ucm234622.htm 42 An SSED is an FDA document (typically a version is submitted by the applicant and modified by FDA) that is intended to present a reasoned, objective, and balanced critique of the scientific evidence that served as the basis of the decision to approve or deny the PMA. The SSED shows that there was reasonable assurance of safety and effectiveness for the device as labeled based on the nonclinical and clinical studies described in the PMA. The SSED is a summation of both the positive and negative aspects of the PMA. For more, see http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/HowtoMarketYourDevice/PremarketSubmission s/PremarketApprovalPMA/ucm050289.htm#ssed. Accessed May 25,2013. 36 The discussion that follows presents the findings in detail. Case examples of certain device types or indications are presented in more detail to highlight when there may be appropriate or insufficient representation of demographic subgroups or underscore unique aspects of certain clinical studies. B. Demographic Subgroup Information Contained in Applications For medical device applications approved in 2011, the working group evaluated the extent to which demographic subgroups, including sex, age, ethnic and racial subgroups, are represented in clinical studies to support medical device approvals and contained in applications. The figures in Part 2 are presented by application (n=33) and exclude applications that reference patient populations that are already represented in another application. 43,44,45 For eight applications, multiple pivotal studies, or cohorts, were submitted for each application, with secondary studies conducted to evaluate a specific subpopulation or device model. Because of the distinct purpose of these studies, information from these multiple studies is presented in a separate section of Part 2. The analysis showed that trial composition varied by product area. In addition to presenting the percentage of representation by each demographic subgroup, the figures below identify the percentage of data not reported for a given study or application. 46 A number of factors influence the interpretation and clinical relevance of demographic information, including, for example, intended population for use, prevalence of disease, and study sample size. 1. Age Composition All 33 approved PMA applications containing unique study datasets provided age composition data. Of the 33 approved PMA applications, age range was reported in 29 applications. Figure 21 depicts the age range and means (when available) of the clinical study populations presented in the PMA applications. Of the 29 PMA applications that reported age range, eight did not report age means. Figure 2-1 excludes four PMA applications that did not report the age range in all clinical studies; when age range was not provided, other age data were included in the submission (e.g., mean, median, standard deviation, number of subjects by age groups, etc.). 43 Some in vitro diagnostic applications were excluded from the demographic data because they contained the same study participants (e.g., different assays were tested using the same patient samples). 44 In one case, an application was submitted for an in vitro diagnostic companion device. No sex, race, or age analyses were conducted for safety and effectiveness endpoints of the in vitro diagnostic companion device. However, sex, race, and age analyses were conducted for the safety and effectiveness of the related drug. 45 Appendix 4 contains a full description of indications associated with each PMA application. 46 The percentage not reported reflects the total proportion of subjects without demographics reported across all the clinical studies included in this analysis for a given PMA application. 37 Figure 2-1: Age Range by Submission (CDRH) The age composition shows that, on average, the study populations for all of the PMAs (reporting a mean age) represented adults. For the clinical studies that reported a mean, the mean age for the study populations ranged from approximately 40 to 75 years. There was a lack of consistency in the type of descriptive statistics on age reported in the PMAs. The data are presented by range in Figure 2-1 because the age range was most commonly reported. The manner in which the age descriptive statistics are presented in a submission may affect data interpretation. FDA provides guidance on developing medical devices for pediatric 38 population subgroups (e.g., neonates, infants, children, and adolescents). 47 However, there is no definition for or device guidance on the geriatric population. FDA currently has several device-related initiatives underway that target the pediatric subpopulation. In 2007, Congress enacted the Pediatric Medical Device Safety and Improvement Act (PMDSIA) as part of the Food and Drug Administration Amendments Act (FDAAA), 48 which provides that FDA may extrapolate adult effectiveness data or performance data that demonstrate probable benefit (in the instance of humanitarian device exemptions (HDEs)) to support a pediatric indication when the course of the disease or condition in children and the effects of the device are similar to adults. FDA is developing a guidance document to describe how effectiveness data on adults can be used for extrapolation to support a pediatric indication. The PMDSIA does not address extrapolation of adult safety data. Additionally, PMDSIA amended the FD&C Act by adding, among other things, a section that requires certain medical device applications to include, if readily available, a description of any pediatric subpopulations that suffer from the disease or condition that the device is intended to treat, diagnose, or cure and the number of affected pediatric patients. 49 In February 2013, FDA published a proposed rule 50 to amend the regulations on premarket approval of medical devices to include requirements relating to the submission of information on pediatric subpopulations that suffer from the disease or condition that a device is intended to treat, diagnose, or cure. In 2011, FDA approved 38 original PMA and panel-track PMA supplements and 3 HDE applications for pediatric use. 2. Sex Composition All of the approved PMA applications evaluated by the working group addressed trial composition by sex. Figure 2-2 shows the results for clinical trial demographic composition by female sex, plotted according to indication or device type. There were five PMA applications with clinical studies representing a 100% female population; these five PMAs were approved for testing or screening 47 FDA Guidance for Industry and FDA Staff on Premarket Assessment of Pediatric Medical Devices. Note: FDA guidance documents are available at http://www.fda.gov/ForIndustry/FDABasicsforIndustry/ucm234622.htm 48 Pediatric Medical Device Safety and Improvement Act (PMDSIA).Public Law No 110-85 (March 13, 2007). Available at http://www.fda.gov/RegulatoryInformation/Legislation/FederalFoodDrugandCosmeticActFDCAct/SignificantAmend mentstotheFDCAct/FoodandDrugAdministrationAmendmentsActof2007/FullTextofFDAAALaw/default.htm. Accessed July 22, 2013 49 See Section 515A (21 U.S.C. 360e-1). Available at http://www.gpo.gov/fdsys/pkg/USCODE-2010title21/html/USCODE-2010-title21-chap9-subchapV-partA-sec360e-1.htm. Accessed May 9, 2013. 50 The proposed rule is available at http://www.gpo.gov/fdsys/pkg/FR-2013-02-19/html/2013-03647.htm. Accessed May 6, 2013. 39 for breast cancer and HPV in females. The PMAs associated with facial wrinkle correction and fecal incontinence devices included approximately 90% females. A higher percentage of women in the facial wrinkle correction clinical study could potentially be due to the aesthetic indication. Fecal incontinence is slightly more common among women because of the effects of pregnancy and childbirth. 51 The PMA for the endovascular occlusion device had 18% female participation, which is not unexpected for this device type. For devices indicated for use in off-pump beating heart surgery, investigators often select patients with larger coronary size and less diffuse nature of coronary disease, which may lead to a skewed distribution of males and females in the trial patient population when compared to the overall sex distribution of all patients with coronary artery disease. Figure 2-2: Sex Composition by Submission (CDRH) 51 National Institutes of Health, 2012. Fecal Incontinence. Publication No. 13-4866, December 2012. Available at http://digestive.niddk.nih.gov/ddiseases/pubs/fecalincontinence/. Accessed May 6, 2013. 40 FDA currently has several device-related efforts ongoing related to sex subgroups. In 2011, FDA issued draft guidance on evaluating sex differences in medical device studies, outlining FDA’s expectations on sex-specific patient enrollment, data analysis, and reporting of study information. 52 The guidance recommends that data from such studies be appropriately analyzed for sex differences. Additionally, in June 2013, CDRH held the Health of Women Public Workshop 53 to discuss how to improve the availability, consistency, and communication of sexspecific information for the safe and effective use of medical devices in women; address identified gaps and unmet needs through targeted resources; and foster the development of innovative strategies, technology and clinical study models. 3. Ethnic and Racial Composition In accordance with FDA’s guidance for industry Collection of Race and Ethnicity Data in Clinical Trials, issued September 2005, 54 patients may self-identify in both an ethnic and racial category (e.g., Hispanic-White, Hispanic-Black). In this FDA guidance, OMB “stated that its race and ethnicity categories were not anthropologic or scientifically based designations, but instead were categories that described the sociocultural construct of our society. The Department of Health and Human Services (HHS) chose to adopt these standardized categories for its agencies that report statistics because the categories are relevant to assessing various health related data, including public health surveillance and research.” FDA accepts applications containing clinical study data with ethnic and racial demographic data captured as one category or separately. In this section, ethnicity and race data are presented in separate plots and discussed individually. • Ethnic Composition Of the 33 approved PMA applications representing unique study datasets evaluated for this report, ethnic demographic data were reported in 23 applications. These data exclude ten PMA applications that did not report any data on ethnic representation. Approximately 46% (21 out of 46) of the pivotal clinical studies that were reviewed collected ethnic demographic data separately from racial data. Figure 2-3 describes the percentage of Hispanics or Latinos by PMA application. Hispanic or Latino representation ranged from 0.3-35%. The findings indicate that PMAs for hepatitis B virus and HPV have the highest percentages of Hispanics. With the exception of the PMAs for lung cancer detection and for the vascular closure device, ethnicity was reported for the majority of clinical trial participants in the study. 52 FDA draft Guidance for Industry and Food and Drug Administration Staff, Evaluation of Sex Differences in Medical Device Clinical Studies. 53 See http://www.fda.gov/MedicalDevices/NewsEvents/WorkshopsConferences/ucm346073.htm 54 FDA guidance Collection of Race and Ethnicity Data in Clinical Trials. Note: FDA guidance documents are available at http://www.fda.gov/ForIndustry/FDABasicsforIndustry/ucm234622.htm. 41 Figure 2-3: Ethnic Composition by Submission (CDRH) Note: the y-axis for this graph only goes up to 40%. • Racial Composition Of the 33 approved PMA applications containing unique study datasets, race was reported in 23 applications. These data exclude 10 PMA applications that did not report any data on racial composition. Two out of the 10 excluded PMA applications were for in vitro diagnostic devices (IVDDs). IVDD PMA applications often involve clinical studies conducted on specimens not directly obtained from patients as part of the clinical trial (e.g., samples obtained from state laboratories or controlled collections) and may lack accompanying demographic information, as compared to samples taken directly at the patient care setting where demographic information can be obtained. Figure 2-4 describes the percentage of each racial subgroup in a PMA application sorted by indication or device type. The racial representation illustrates that, in most PMAs, there was a predominant representation of the White population. For three PMAs (hepatitis B virus 2, 42 hepatitis B virus 3 and facial wrinkle correction device) a relatively high percentage of African Americans or Blacks compared to the other racial subgroups were included. For the facial wrinkle correction device, one of the clinical studies conducted for this PMA assessed the safety and effectiveness of the device specifically in non-White populations. For some of the indications, racial composition in the trials was consistent with the prevalence of the disease across different racial subgroups when adjusted by the overall population distribution of racial subgroups. For example, African Americans or Blacks have a higher prevalence of infection with hepatitis B 55 and were similarly observed to have a higher participation in the hepatitis B clinical studies compared to other racial subgroups. For other indications, racial composition did not appear to represent disease prevalence in the U.S. population. For example, although African Americans have the highest rates for lung cancer 56 and chronic hepatitis C infection, 57 this was not reflected in the percentage of African Americans enrolled relative to Whites and Asians/Native Hawaiians/Other Pacific Islanders in the clinical studies for these diseases. In some cases, this could be attributable to the inclusion of foreign study sites (e.g., Asia) in the clinical studies, where African Americans would be less likely to be enrolled. 55 Centers for Disease Control and Prevention. 2012. Health Disparities in HIV/AIDS, Viral Hepatitis, STDs, and TB. July 11, 2012. Available at http://www.cdc.gov/nchhstp/healthdisparities/AfricanAmericans.html#Hepatitis. Accessed May 6, 2013. 56 Centers for Disease Control and Prevention. Lung Cancer Rates by Race and Ethnicity. December 19, 2012. Available at http://www.cdc.gov/cancer/lung/statistics/race.htm. Accessed May 6, 2013. 57 Centers for Disease Control and Prevention. Hepatitis C in the African American Community. May 21, 2012. Available at http://www.cdc.gov/hepatitis/populations/AAC-HepC.htm. Accessed May 6, 2013. 43 Figure 2-4: Race Composition by Submission (CDRH) Abbreviations: NH = Native Hawaiian, OPI = Other Pacific Islander, AI = American Indian, AN =Alaska Native C. Multiple Pivotal Studies Of the 37 PMAs approved in 2011, 8 were supported by 2 pivotal studies or cohorts. These second pivotal studies were often conducted to evaluate the medical device or diagnostic in a specific subgroup. Consequently, the subgroup demographics of the eight PMAs with multiple studies were assessed separately. Several of the cardiovascular PMAs evaluated the device in the primary study cohort and in patients with smaller blood vessels. In a second example, a facial wrinkle correction device was supported by one study targeting the primary study cohort while another study specifically targeted patients with darker skin pigmentation. Additionally, a product to detect antibodies to the hepatitis e-antigen was evaluated in both the primary and pediatric-study cohorts. For all eight PMAs with multiple pivotal study cohorts, trial composition was reported for all demographic categories for at least one of the study groups. As previously stated, multiple 44 factors influence the interpretation and clinical relevance of demographic information, such as intended population for use, prevalence of disease, and study sample size. 1. Age Composition Figure 2-5 describes the age range and mean age (if available) of the study populations presented by PMA application for PMAs with multiple clinical studies. The results of this age analysis show that, on average, the study populations for PMAs with multiple pivotal studies (where mean age was reported) were adults. For the pivotal studies that reported a mean, the mean ranged from 35 to 65 years old. Figure 2-5 highlights the pediatric subpopulation presented in the hepatitis B virus 2 PMA. As described earlier, inconsistent descriptive statistics were provided for age data in these PMAs. The data are presented by range because the PMAs most commonly reported the age range. 45 Figure 2-5: Age Range by Study for Submissions with Two Studies (CDRH) 2. Sex Composition Figure 2-6 describes the female sex composition data presented by study for PMAs with two pivotal studies. As described previously, the facial wrinkle correction PMA consisted of two pivotal studies evaluating the safety and effectiveness of the device in the primary study cohort and in a specific racial subgroup. Both pivotal studies for this PMA had a high representation of females (approximately 85 and 95%). The HPV clinical studies enrolled only females. 46 Figure 2-6: Sex composition by Study for Submissions with Two Studies (CDRH) 3. Ethnic and Racial Composition In this section, race and ethnicity data are presented in separate plots and discussed individually. • Ethnic Composition Figure 2-7 describes the percentage of Hispanics or Latinos by PMA for PMA applications with multiple pivotal studies. For two PMAs (coronary drug eluting stent 1 and vascular closure device), the ethnic composition for one of the study cohorts was not reported. There was generally a low representation of Hispanics across the eight PMAs with multiple studies, with the exception of the hepatitis B virus 2 PMA, in which the pediatric clinical study had more than 40% Hispanic representation (see Case Example), and both HPV PMAs, which had Hispanic representation ranging from approximately 20–35%. Two pivotal studies were conducted for both HPV PMAs because of the indications specifically sought/approved in the PMAs (i.e., one clinical study enrolled women ≥21 years of age while the other clinical study enrolled women ≥30 years of age). 47 Case Example: Hepatitis B Virus The hepatitis B virus 2 PMA involved two pivotal clinical study populations: one targeting the primary study cohort, the other a pediatric cohort. The pediatric clinical study consisted of samples from a population of pediatric patients in Florida at high risk for exposure to viral hepatitis. Demographic characteristics of the pediatric clinical study (n=165) showed that Hispanics represented 42.4% of the study cohort. The higher Hispanic participation in the pediatric study in comparison to the primary study cohort could be explained by the enrollment of patients from a single geographic location where high Hispanic enrollment would be expected. This example illustrates that for studies with few enrollment sites, the demographics of the site may have a substantial effect on the distribution of patients enrolled. Figure 2-7: Ethnic Composition by Study for Submissions with Two Studies (CDRH) Abbreviations: N/R = Not reported Note: The y-axis for this graph only goes up to 40%. 48 • Racial Composition With the exception of the facial wrinkle correction (see Case Example) and the hepatitis B virus 2 PMAs, the racial composition for all of the PMAs with multiple pivotal studies was predominantly White. For two PMAs (vascular closure device and coronary drug eluting stent 1), the racial composition for one of the study cohorts was not reported. Figure 2-8 depicts the racial composition data presented by study for PMAs with two clinical studies. Case Example: Facial Wrinkle Correction PMA The facial wrinkle correction device was evaluated in clinical studies representing the primary study cohort and in patients with Fitzpatrick Skin Phototype Scores ≥ IV. The primary study (n=118) enrolled a predominantly White (96.6%) population. Minority populations made up less than 4% of the study population. The sponsor conducted the Fitzpatrick premarket study to better understand the safety and effectiveness of the facial wrinkle correction device in persons of color. The Fitzpatrick clinical study (n=93) enrolled predominantly African American or Black subjects (96.8%). Figure 2-8: Racial Composition by Study for Submissions with Two Studies (CDRH) Abbreviations: N/R = Not reported, NH=Native Hawaiian, OPI=Other Pacific Islander, AI=American Indian, AN=Alaska Native 49 D. Subset Analysis in Applications and Availability of Summary Information For medical device applications approved in 2011, the working group evaluated the extent to which demographic subset analyses on sex, age, race, and ethnicity were included in PMA applications and summary information was publicly available. Public documents on PMA approvals and evidence of demographic subgroup analyses include the product labeling and the SSED, 58 which are both posted on FDA’s Website, shortly after PMA approval. FDA does not have standard labeling content requirements for devices (apart from the requirements for in vitro diagnostic devices). This results in inconsistent reporting of this information in device labeling. The figures in this section are presented by PMA (n=32 for sex analyses and n=37 for race/ethnicity and age analyses) and include PMAs using the same patient population as another PMA application. Five PMA applications approved for screening for breast cancer or HPV were excluded from the sex analysis because these PMAs enrolled a female-only study population. Figure 2-9 describes the percentage of PMA applications containing demographic subgroup analyses. Results show that 88% of the PMA applications reviewed contained a sex analysis; 70% contained an age analysis; and 27% contained race or ethnicity analyses. Because only two of the 37 PMAs contained ethnicity subgroup analyses, race and ethnicity were grouped together for this assessment. Figure 2-10 describes the public availability of summary information (in the device labeling and/or SSED) regarding demographic subset analyses. Results show that 63% of the PMA applications reviewed had a statement conveyed in the device labeling and/or SSED on sex subgroup analysis; 57% had a statement related to an age analysis; and 16% had a statement related to a race or ethnicity analysis. This demonstrates that FDA publicly communicated information on subgroup analyses for sex and age for more than 50% of the PMA applications approved in 2011. Race or ethnicity analyses were submitted in PMA applications and conveyed to the public to a lesser extent. The findings of the working group highlight that to a large extent demographic subset analyses are present in PMA applications, but FDA’s communication to the public of the results of those demographic subset analyses is less consistent. 58 An SSED is an FDA document (typically a version is submitted by the applicant and modified by FDA) that is intended to present a reasoned, objective, and balanced critique of the scientific evidence that served as the basis of the decision to approve or deny the PMA. The SSED shows that there was reasonable assurance of safety and effectiveness for the device as labeled based on the nonclinical and clinical studies described in the PMA. The SSED is a summation of both the positive and negative aspects of the PMA. For more, see http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/HowtoMarketYourDevice/PremarketSubmissio ns/PremarketApprovalPMA/ucm050289.htm#ssed. Accessed May 25, 2013. 50 Figure 2-9: Presence of Subgroup Analysis in Applications (CDRH) Figure 2-10: Presence of Public Statement about Subgroup Analysis (CDRH) * Excludes five applications with breast cancer or human papillomavirus indications Figure 2-11 describes the percentage of PMA applications that publicly convey (in device labeling and/or SSED) summary information on demographic subset analyses, indicated by whether data or conclusions were presented or whether only a statement was included regarding lack of conclusions from subset analyses. For sex and age subset analyses, there was a higher proportion of PMAs that presented data or conclusions compared to PMAs that included only a statement about lack of conclusions regarding demographic subset analyses. For race or ethnicity subset analyses, the same proportion of PMAs had data or conclusions or a statement about lack of conclusions regarding subset analyses. 51 Figure 2-11: Presence of Public Statement about Subgroup Analysis (CDRH) * Excludes five applications with breast cancer or human papillomavirus indications E. Summary of Findings for Part 2: Medical Devices In summary, the general findings based on this 2011 sampling of medical devices and diagnostics reviewed for this report include the following: • All PMA applications reported trial composition by sex and age, and the majority of the submissions reported race and ethnicity composition. Demographic subset analyses were more commonly available for sex and age than for race or ethnicity. • FDA considered demographic subset analyses in PMA applications more often than was presented in device labeling and FDA communications. Moreover, nearly one quarter of 52 the PMAs contained a second pivotal study cohort, which was designed to obtain clinical experience in a specific subpopulation. 59 • The representation of demographic subgroups varied widely by product area. This is because a number of factors can influence the interpretation and clinical relevance of demographic information (e.g., intended population for use, prevalence of disease, study sample size). • Although it is important to include diverse populations in clinical trials when possible and appropriate, the unique nature of medical devices means this additional information would not always be contributory to FDA’s decision making. For example, certain microbiology in-vitro diagnostic devices (IVDDs) have high overall accuracy; when this occurs, additional analyses by subpopulations would not affect clearance or approval. FDA Tools for Communicating Demographic Information When drugs, biologics, and devices are approved for marketing in the United States, FDA approved labeling is made publicly available. Thus, all FDA-approved products have public information about safety and effectiveness available at the time of marketing. As required under Section 916 of the Food and Drug Administration Amendments Act of 2007 (FDAAA), 60 FDA communicates information about known differences in safety and effectiveness of drugs and biologics by demographic subgroup to the public through a variety of product approval-related documents posted on its Web site. For example, for drugs and biologics, clinical and statistical reviews and product labeling, in addition to other review-related documents, are typically posted on FDA’s Web site. Action packages (a comprehensive collection of FDA staff reviews, labeling, selected correspondence, and other relevant documents for an application) for drugs and biologics are required to be posted to FDA’s Web site within 30 calendar days of approval for new products or within 30 calendar days of the third Freedom of Information Act (FOIA) 61 request for the action package. 62 A summary review (summary basis of regulatory action 59 Three PMAs evaluated diagnostic devices in specific age subpopulations; one PMA evaluated an aesthetic device to a specific race subpopulation; and four PMAs evaluated cardiovascular devices specifically in patients with smaller blood vessel sizes. 60 Food and Drug Administration Amendments Act of 2007. Public Law No.110-85 (September 27, 2007). Available at http://www.fda.gov/RegulatoryInformation/Legislation/FederalFoodDrugandCosmeticActFDCAct/SignificantAmen dmentstotheFDCAct/FoodandDrugAdministrationAmendmentsActof2007/FullTextofFDAAALaw/default.htm. Accessed April 4, 2013. 61 Freedom of Information Act (FOIA). Public Law No 89-487 (July 4, 1966). Available at http://www.foia.gov/. Accessed April 4, 2013. 62 Standard Operating Policies and Procedures (SOPP) 8401.7 Action Package for Posting. Available at http://www.fda.gov/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/ProceduresSOPPs/ucm21 1616.htm. Accessed June 13, 2013. 53 (SBRA) is required to be posted within 48 hours of approval unless redaction is required. 63 For devices, the SSED along with the approval order, device labeling, and other consumer information (e.g., brief overview of the product) are made publicly available. These materials are posted on the Web site shortly after product approval. 64 A number of FDA-wide posting and disclosure policies apply as well. These policies are set based on legal mandates and technical requirements that must be followed to successfully comply with FDAAA, the Americans with Disabilities Act of 1990 (ADA) 65 for accessibility under Section 508 of the Rehabilitation Act of 1973, 66 and other federal statutory requirements. Tools Related to Pediatric Demographics FDAAA requires that FDA track and make publicly available certain pediatric information from pediatric clinical trials for all medical products. 67 Additionally, under PREA, pediatric-focused reviews and labeling changes are posted on a dedicated page on FDA’s Web site. 68 In accordance with FDAAA, FDA’s Web site provides information from pediatric studies conducted in response to Written Requests (per Section 505A of BPCA) and pediatric assessments (per PREA). Unless otherwise noted, these reports contain information from both CBER and CDER. On February 19, 2013, FDA issued a proposed rule that requires manufactures to submit publicly available information on pediatric patients (21 years or younger) that suffer from the disease or condition that the devices submitted for FDA approval are intended to treat, diagnose, or cure. This information will help key stakeholders, device manufacturers, and the FDA track what devices are available to pediatric patients and identify unmet pediatric device needs. Communication of Clinically Significant Findings in Demographic Subgroups In general, for all medical products, if a review demonstrates that a clinically significant difference, or a notable lack of information, exists in safety or efficacy data for a demographic subgroup, the appropriate document (e.g., clinical review, labeling) will reflect those findings. If 63 Ibid. 64 See CDRH Web site. Available at: http://www.fda.gov/MedicalDevices/default.htm. Accessed May 9, 2013. 65 Americans with Disabilities Act of 1990. Public Law 110-325 (January 1, 2009). Available at http://www.ada.gov/pubs/adastatute08.pdf. Accessed March 7, 2013. 66 The Rehabilitation Act Amendments of 1973. Public Law No 93-112 (September 26, 1973). Available at http://www.access-board.gov/enforcement/rehab-act-text/intro.htm. Accessed March 7, 2013. 67 Food and Drug Administration Amendments Act of 2007. Public Law No.110-85 (September 27, 2007). Available at http://www.fda.gov/RegulatoryInformation/Legislation/FederalFoodDrugandCosmeticActFDCAct/SignificantAmen dmentstotheFDCAct/FoodandDrugAdministrationAmendmentsActof2007/FullTextofFDAAALaw/default.htm. Accessed April 4, 2013. 68 See for example, FDA—Biologics PREA Reviews and Labeling Changes. Available at http://www.fda.gov/AboutFDA/CentersOffices/OfficeofMedicalProductsandTobacco/CBER/ucm122938.htm. Accessed April 4, 2013. 54 significant safety concerns exist for certain subpopulations, this information is provided in the appropriate section of the product labeling (e.g., Warning & Precautions, Pediatric Use, Geriatric Use). Products labeled for pediatric use, are also posted according to the Pediatric Tracking Requirements under FDAAA. 69 Upon approval of a new drug or biologic for marketing, FDA posts the following documents (as applicable) on the relevant Web site 70 (i.e., the Action Package): • • • • • • • Labeling information Approval history Certain letters (to the company) Reviews and related documents (including action packages) Risk evaluation and mitigation strategy (REMS) Medication guides Other important information from FDA The timeliness with which approved medical product information is posted depends on whether the documents must undergo pre-posting redaction according to the regulations pertinent to confidential information. In general, letters and review packages do require redaction, but labels do not. When redaction is needed, the posting schedule for approval letters and labeling is typically one-to-two days after approval (the approval letter is where information about any postmarket commitments and requirements for the drug or biologic can be found). Action packages take longer to process because of their size and complexity and the amount of time needed prior to posting can range from one week to six months. For devices, the SSED along with the approval order, device labeling, and other consumer information (e.g., brief overview of the product) are made publicly available. These materials are posted on the Web site generally within a few days of approval. 71 FDA Advisory Committee Meetings In addition, many FDA products are routinely brought before FDA Advisory Committees during which a medical product’s safety and/or effectiveness is discussed. These discussions may include risk–benefit assessments in special populations and presentations from interested 69 For the purposes of this report, pediatric data are reported if the primary indication is for pediatrics. However, in general, consistent with the Pediatric Tracking Requirements under the Food and Drug Administration Amendments Act (FDAAA), pursuant to the Best Pharmaceuticals for Children Act and the Pediatric Research Equity Act, pediatric demographic data for labeling changes are collected in the Pediatric Study Characteristics Database, which is managed by FDA’s Office of Pediatric Therapeutics (OPT). In collaboration with FDA’s Centers, OPT collects pediatric study data and posts them to FDA’s Web site. 70 For medical products approved in CDER, see Drugs@FDA. Available at http://www.accessdata.fda.gov/scripts/cder/drugsatfda/. Accessed May 9, 2013. For products approved in CBER, see Vaccines, Blood & Biologics. Available at www.fda.gov/BiologicsBloodVaccines/default.htm. Accessed April 4, 2013. 55 stakeholder groups. Advisory Committee meetings are public meetings, and have background packages and other materials, including full meeting transcripts, posted on FDA’s Web site. Communication of Postmarket Safety Concerns For all FDA-regulated medical products, if safety concerns are identified in the postmarket period (i.e., after products are approved for marketing), FDA will ensure that appropriate information is communicated to health care professionals and the public through a variety of methods, including the Medwatch safety alert system, 72 press releases and special communications, 73 FDA’s social media sites 74 (Facebook and Twitter accounts), and communication tools that FDA can access when necessary through a variety of partnerships with health care professional and public health associations. 75 Public Workshops FDA periodically holds workshops and public meetings on demographic subgroup inclusion and analyses; some examples follow. • In 1995 a workshop titled Gender Studies in Product Development: Scientific Issues and Approaches 76 explored the science involved with assessing gender effects during development of medical products, including drugs, biologics, and medical devices, and identified significant areas for further research and policy development. • In 2006, an important workshop titled, Sex Differences and the FDA Critical Path Initiative 77 addressed the importance of understanding the biological differences between men and women in the context of developing tools to improve and accelerate the development of drugs, biologics, and devices. 72 Safety—MedWatch Safety Alerts for Human Medical products. Available at http://www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsforHumanMedicalProducts/default.htm Accessed April 4, 2013. 73 Drugs—Drug Safety and Availability. Available at http://www.fda.gov/Drugs/DrugSafety/default.htm. Accessed April 4, 2013. 74 News & Events—Interactive Media. Available at http://www.fda.gov/NewsEvents/InteractiveMedia/default.htm#Facebook. Accessed April 4, 2013. 75 FDA guidance Drug Safety Information – FDA’s Communication to the Public. Note: FDA guidance documents are available at http://www.fda.gov/ForIndustry/FDABasicsforIndustry/ucm234622.htm. 76 Food and Drug Administration Office of Women's Health. 1995. Gender Studies in Product Development: Scientific Issues and Approaches. Available at http://www.fda.gov/ScienceResearch/SpecialTopics/WomensHealthResearch/ucm134457.htm. Accessed May 9, 2013. 77 See http://www.womenshealthresearch.org/site/DocServer/FDA_Critical_Path.pdf?docID=1401. Accessed May 9, 2013. 56 • In 2008, CDRH held two public workshops to discuss ways to overcome barriers to understanding the effect of sex differences on clinical outcomes, with a focus on clinical study conduct and statistical analysis: Exploration of Public Policy Development Regarding the Study and Analysis of Sex Differences in the Clinical Evaluation of Cardiovascular Medical Products 78 and FDA/Advanced Medical Technology Association (AdvaMed) Workshop: Gender Differences in Cardiovascular Devices. 79 • In 2011, a conference titled Dialogues on Diversifying Clinical Trials: Successful Strategies for Engaging Women and Minorities 80 (cosponsored by the Society for Women’s Health Research) focused on novel methods for improving recruitment and retention of women and minorities, community-based approaches to clinical trial design, federal perspectives on guidelines and regulations to improve diversity in government- and industry-funded research. • In June 2013, CDRH held the Health of Women Public Workshop 81 to discuss how to improve the availability, consistency and communication of sex-specific information for the safe and effective use of medical devices in women, address identified gaps and unmet needs through targeted resources and foster the development of innovative strategies, technology and clinical study models. Scientific Publications In addition, scientific publications by FDA review scientists can be a source of relevant demographics information for approved drugs, biologics, and medical devices (see selected examples of relevant FDA publications in Appendix 5). 78 See http://www.medicaldevicestoday.com/2008/06/cdrh-explores-r.html. Accessed May 9, 2013. 79 See http://www.gpo.gov/fdsys/pkg/FR-2008-11-26/html/E8-28169.htm. Accessed May 9, 2013. 80 See http://www.fda.gov/ScienceResearch/SpecialTopics/WomensHealthResearch/ucm273816.htm. Accessed May 9, 2013. 81 See http://www.fda.gov/MedicalDevices/NewsEvents/WorkshopsConferences/ucm346073.htm. 57 Summary and Conclusions FDA has a variety of statutory, regulatory, and policy-related tools that provide a framework for guiding medical product sponsors on the collection and subset analysis of demographic data on the participants in their clinical trials. FDA’s internal policies and procedures and regulations facilitate the assessment of demographic subgroup information included in marketing applications. Moreover, following medical product approval, FDA communicates available information to the public on the demographic profile of the study participants and on the demographic data subset analyses using a variety of mechanisms (e.g., consumer updates, safety alerts, label changes, etc.). Our analysis for this report determined that, in general, medical product developers/manufacturers and FDA staff are complying with relevant requirements and guidance. Nevertheless, areas for improvement have been noted, and these findings will be used to inform and guide development of the FDASIA Section 907-required Action Plan. Congress asked us to consider four topic areas in preparing this report. The highlights from our study include the following: Tools to ensure submission of demographic information • Although there is some variation by product area, FDA’s statutory and regulatory requirements, guidances, policies, and procedures generally inform sponsors about including tabulations of the demographic data of clinical trial participants and demographic subset analyses in their medical product applications (see Appendix 1). • Similarly, tools (e.g., application review templates and FDA standard operating policies and procedures) guide regulatory review staff in the assessment of marketing applications to ensure that demographic data and subset analyses are included in the information FDA uses in its review and approval processes. Extent of demographic subset analyses • The extent to which demographic subset data were analyzed varied across medical product types (drugs, biologics, and devices). Applications for drugs and biologics addressed subset analyses by sex, race, and age ― that is, the applications mentioned demographic subsets in some way. The majority of the device applications contained a subset analysis for age and sex, with a lower percentage of applications containing a subset analysis for race or ethnicity. Inclusion did not necessarily mean that the data on patient subgroups was sufficient for meaningful analysis or to detect relevant subgroup effects. In some of the applications reviewed for this report, the results of the subgroup analyses were limited by low sample size. 58 Extent of demographic subgroup representation in clinical trials • • • All biologics, drugs, and the majority of the medical device applications reviewed for this report provided the composition of clinical study participants by age, race, and sex. Participants’ sex was the most consistently reported in the medical product applications. For approved drugs and biologics, the extent to which patients were represented in clinical trials by age and sex tended to reflect the disease indication studied. For devices, patient participation by age and sex varied by product area. Although it is important to include diverse populations in clinical trials when possible and appropriate, the unique nature of medical devices means this additional information would not always be contributory to FDA’s decision making. For example, certain microbiology in-vitro diagnostic devices (IVDDs) have high overall accuracy; when this occurs, additional analyses by subpopulations would not impact clearance or approval. Whites represented a high percentage of clinical trial study participants for biologic, drug, and medical device applications. In many cases, other racial subgroups were underrepresented. Communication of demographic subgroup information to the public • FDA’s internal policies and procedures and regulations facilitate the assessment of demographic subgroup information included in marketing applications. Moreover, following medical product approval, FDA can communicate available information to the public on the demographic profile of the study participants and on the demographic data subset analyses using a variety of mechanisms: initially with product labeling and publicly posted clinical reviews and later, once the product is on the market, with consumer updates, safety alerts, label changes, and other mechanisms, should this be necessary. • Statutory differences in the regulatory framework for medical devices compared to those applicable to drugs and biologics account for differences in policies and practices across FDA centers with regard to submission and analysis of demographic data and public disclosure of information at the time a product is approved (e.g., timing of information, information release, and public documents). There are a number of other important issues to consider when interpreting the results in this report. Apparent differences among demographic groups that can affect health-related behaviors and health outcomes can be influenced by two broad categories of factors that often interact and overlap: (1) extrinsic factors (e.g., socioeconomic and cultural influences, diet, environment); and (2) intrinsic biological factors (e.g., genetics, hormones, metabolism, organ function, body weight). 82 Inclusion of broad demographic subgroups (such as diverse racial and ethnic groups) 82 FDA guidance Collection of Race and Ethnicity Data in Clinical Trials. Note: FDA guidance documents are available at http://www.fda.gov/ForIndustry/FDABasicsforIndustry/ucm234622.htm. 59 in clinical trials has the potential to directly address the first category. However, achieving diverse demographic subgroup participation in clinical trials remains a challenge after decades of efforts on the part of the broad stakeholder community, including patients and advocates, industry, academia, and public health agencies. These challenges involve a complex interplay of socioeconomic factors, such as practical considerations for participating in clinical trials (e.g., location of study centers). As we move into the coming decades, FDA’s regulatory mission will increasingly focus on gathering and understanding information related to the second category, intrinsic factors— genetic and biological influences that affect disease and response to medical products (effectiveness and safety). One of FDA’s goals is to make regulatory decisions based on scientific information and to publicly communicate actionable information. That is, when clinically meaningful differences are observed for certain subgroups (e.g., an adverse effect seen more commonly with a certain genetic mutation), this information is included in the product labeling or otherwise publicly released. This information is then used to guide health care professionals in prescribing and monitoring products used by their patients. The science of variability in human response to medical treatments is driven by many factors, which science is continuing to address over time. In many cases, when demographic subgroup data are collected and analyzed, the response to a drug, a biologic, or a device is generally similar across demographic groups. There are exceptions when clinically meaningful differences in response to a medical product have been observed in certain subgroups of the population. In some cases, genetic or other biological factors may drive variability. In other cases, observed differences may be attributable to other intrinsic or extrinsic factors, or interactions between these or other factors. For these reasons, there are limitations to focusing on demographic analysis without the context of other factors driving variability. Although it is important to include diverse populations in clinical trials whenever possible and appropriate, the broad self-identified demographic categories used today may not relate to the complex genetic and biological factors that are the basis for differences in response to medical products, 83,84 although they may be useful in generating hypotheses that may drive additional studies or product development in the future. But in many situations, the demographic subset analysis provided by sponsors may not be actionable. For example, vaccines are generally highly effective in all demographic subgroups. In these cases, FDA has not included information for individual demographic groups in public documents. It is also important to note that evidence related to the safety or effectiveness of a medical product is based not only on the data from clinical trials, but also on data that may come from a variety of sources once a product is available on the market (e.g., postmarket surveillance). When a product goes on the market, a much broader population uses the product, and if new data 83 Rotimi CN, Jorde LB. Ancestry and disease in the age of genomic medicine. New Engl J Med 2010;363:1551-8. 84 Jaimes N, Londono V, Halpern AC. The term Hispanic/Latino. A note of caution. JAMA Dermatol 2013;149:274-5. 60 suggest an actionable difference in certain groups of patients, FDA communicates this information to the public using a variety of methods. Our increasing understanding of intrinsic biological factors is improving our ability to tailor treatments at the level of the individual (i.e., personalized medicine). Numerous FDA initiatives have been launched to advance the promise of personalized medicine. 85 These initiatives seek to determine what is the correct medical intervention at the correct time for each patient. This targeted approach is becoming the focus of modern medical product development as scientific advancements delineate the specific genetic variables underlying many diseases. FDA recognizes the importance in its public health mission of effectively communicating clinically meaningful differences that are observed for certain subgroups, including demographic subgroups, to inform decisions by health care practitioners and patients. To that end, FDA has initiated efforts to develop and publicly release an Action Plan for strengthening the availability of such data, as required in FDASIA Section 907. To support that effort and leverage relevant stakeholder expertise, FDA has opened a docket 86 in connection with the release of this report to solicit input and recommendations from the public. 85 Drugs, Genomics. FDA Web site.